A triple-slot active reflectarray cell using a ferroelectric capacitor

HAL Id: hal-01116821

https://hal.archives-ouvertes.fr/hal-01116821

Submitted on 16 Feb 2015

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

A triple-slot active reflectarray cell using a ferroelectric capacitor

Kevin Nadaud, Raphaël Gillard, Erwan Fourn, Caroline Borderon, Hartmut Gundel

To cite this version:

Kevin Nadaud, Raphaël Gillard, Erwan Fourn, Caroline Borderon, Hartmut Gundel. A triple-slot

active reflectarray cell using a ferroelectric capacitor. European Conference on Antennas and Propa-

gation, Apr 2015, Lisbonnes, Portugal. �hal-01116821�

A triple-slot active reflectarray cell using a ferroelectric capacitor

Kevin Nadaud

, Raphaël Gillard

, Erwan Fourn

, Caroline Borderon

and Hartmut W. Gundel

1

IETR, University of Nantes, Nantes, France

2

IETR, INSA of Rennes, Rennes, France

kevin.nadaud@univ-nantes.fr, raphael.gillard@insa-rennes.fr

Abstract—A tunable reflectarray phase-shifting cell, designed for a resonance frequency of 5.5 GHz, is presented. The cell is based on a triple-slot topology and loaded by a ferroelectric thin film capacitor with 60% tunability under 400 kV/cm bias electric field. The use of three slots allows reducing the losses and increasing the bandwidth of the cell. The cell provides 270 degrees of phase range with a maximum of 6 dB loss in the band.

Index Terms—Ferroelectric, reflectarray, thin film, tunability.

I. I

Planar reflectarrays are widely studied because they com- bine the advantages of printed array antennas and reflector antennas [1]. Active cells enable reconfigurable reflectarrays whose radiating beam can be steered or shaped dynamically.

Active cells usually derive from passive topologies in which electronic devices are integrated in order to control the desired phase-shift [2], [3]. Advanced tunable materials like Liquid Crystals [4] or ferroelectrics [5] can also be used. Ferroelectric materials have the advantage of no bias current, which may considerably reduce power consumption.

Most passive and active cells are based on resonance phenomena to obtain the appropriate phase range. Those kinds of cells may exhibit sharp resonances which cause a narrow bandwidth and significant losses. To reduce the losses and increase the bandwidth, multiple-resonator cells have been used with various topologies such as stacked patches [6], coupled dipoles [4] or coupled slots [7].

In [8], the principle of a ferroelectric-loaded slot for the realization of a reconfigurable cell has been demonstrated.

However, this cell exhibits significant losses due to its sharp resonance. An improved cell using three slots with a single capacitive load has been proposed in [7] and validated by simulation. In the present paper we propose to assess its per- formance when a realistic technological process is considered for achieving the tunable capacitive loading.

II. F

M

Ferroelectrics are non-linear materials well known for their high and tunable relative permittivity. They can be used in different microwave devices such as phase-shifters, filters, antennas or reflectarray cells. In the proposed device, Barium- Strontium-Titanate (BST) has been chosen because of its relatively low dielectric losses and high tunability. Moreover BST has a low coercive field, which considerably reduces

L

L

W

W

e

W

m W

L

W

S

S

W

W

S

Active area

1 mm

1 mm

x y

(a)

PCB Alumina

Copper alumina Copper

on Rogers

Bounding copper BST

Waveguide

Ground d

(b)

Figure 1. Topology of the proposed cell and zoom on the capacitor region (top view) (a). The dielectric stack in the waveguide (side view) (b).Ls1= 8 mm ; Ls2 = 18 mm ;Ws1 = Ws2 = 5 mm ;e_f = 3.5 mm ; Wc = 4 mm ; Lc= 3 mm ;We = 30 µm ;m= 2.5 mm ;Wi = 1 mm ;Si = 0.4 mm ; Se=Wcond = 30 µm ;Wcapa = 190 µm ;Scapa = 7 µm ;d= 15.7 mm.

the hysteresis effect and thus simplifies the electric command.

The films were elaborated on alumina substrates by chemical solution deposition [9]. A one-percent manganese-doping has been used in order to reduce low frequency leakage currents and to prevent from breakdown while biasing the material [10].

Dielectric characterization has been performed using a tech- nique described in [11]. At 5 GHz, the thin films have a relative permittivity of 350, a tunability of 60% under 400 kV/cm and a tan δ = 0.02. Details of the material characterization shall be published elsewhere.

III. C

D

The proposed unit-cell consists of three slots in a ground

plane as shown in Fig. 1a and is designed for operation at

5.5 GHz. Both external slots have the same length in order to

preserve the cell symmetry. The presence of these slots allows

reducing the frequency dispersion and losses by providing a

second resonance. As demonstrated in [7], the electrical length

R ef le ct ed p ha se ( °)

−540

−360

−180 0 180

Ebias = 0 kV/cm Ebias = 100 kV/cm Ebias = 200 kV/cm Ebias = 400 kV/cm

R ef le ct ed m ag ni tu de ( dB )

−15

−10

−5 0

Frequency f (GHz)

5 5.2 5.4 5.6 5.8 6

Figure 2. Simulated reflection coefficient (phase and magnitude) of the proposed cell for different applied bias field.

of the central slot controls the phase-shift of the reflected wave. By loading this slot with a tunable capacitor, it is possible to change its electrical length and thus the phase-shift dynamically.

For the sake of simplicity (regarding measurement), the cell is embedded in a 35×35 mm

square metallic waveguide.

Fig. 1b shows the details of the used dielectric stack. The alumina substrate (25.4×25.4 mm

, thickness 508 µm and ε

= 9.8) with the integrated BST thin film is reported on a printed circuit board (PCB, ε

= 2.17, tan δ = 9.10

and thickness 1.6 mm) which is mounted 15.7 mm above the ground plane ending the metallic waveguide. In order to reduce the dielectric losses, the ferroelectric material has been limited by wet etching to only a 1×1 mm

surface for hosting the capacitor. The thickness of the ferroelectric film is 1 µm. The electrode metallization (1.4 µm thick) is realized by sputtering copper and patterning with a classical photolithography process.

The widths and the lengths of the slots have been optimized for obtaining as flat as possible phase states (in order to reduce losses) and a sufficient phase range with available tunability of the BST film. The distance between the slots e

controls the mutual coupling and has been chosen to minimize the phase slope while preserving enough space for the biasing circuit.

The tunable capacitor consists of two coplanar electrodes deposed on the ferroelectric film which are separated by a 7 µm wide gap S

(zoom of Fig. 1a). The biasing circuit can be seen in Fig. 1a. Isolation of the biased electrode from

Phase standard deviation (°)

25 30 35 40 45 50

Frequency f (GHz)

5 5.2 5.4 5.6 5.8 6

10-30 20 40 15-30 15-35

Figure 3. Phase standard deviation computed from 4 chosen states of the proposed cell.

Cross polarization (dB)

−50

−40

−30

−20

−10 0

Frequency f (GHz)

5 5.2 5.4 5.6 5.8 6

Ebias = 0 kV/cm Ebias = 100 kV/cm Ebias = 200 kV/cm Ebias = 400 kV/cm

Figure 4. Simulated cross polarization of the proposed cell for different applied bias field.

the ground is done by a narrow gap (S

width). In addition, a low-pass distributed filter is necessary for improving the decoupling between the DC bias and the RF signal. It consists of an inductive line (width W

and length L

) followed by a capacitive line (width W

and length L

).

IV. S

The simulations of a unit cell have been performed using the waveguide approach with HFSS

commercial software.

The section of the waveguide is 35×35mm

and the excited mode is the TE

mode (according to Fig. 1a). The relative permittivity of the active material has been varied from 350 to 175, which corresponds to a maximum applied bias field of 400 kV/cm. The different simulations are reported in Fig. 2.

The use of multiple resonators is clearly visible because the phase-shift has a quite smooth evolution versus frequency and the phase-shift is provided in a wide band, contrary to what is obtained with single-resonator cells [8]. The losses are still quite high (up to 6 dB) but lower than those reported in [5], [8] for similar materials.

The bandwidth of the cell has been estimated by choosing

4 states (corresponding to 4 values of the electric bias field)

and computing the standard phase deviation [12]. The obtained

Reflected phase (°)

−540

−360

−180 0 180

Scapa = 6 µm Scapa = 7 µm Scapa = 8 µm Ebias = 400 kV/cm

Ebias = 0 kV/cm

Ebias = 400 kV/cm

Ebias = 0 kV/cm

Reflected magnitude (dB)

−15

−10

−5 0

Frequency f (GHz)

5 5.2 5.4 5.6 5.8 6

(a)

Reflected phase (°)

−540

−360

−180 0 180

hf = 0.9 µm hf = 1 µm hf = 1.1 µm Ebias = 400 kV/cm

Ebias = 0 kV/cm

Ebias = 400 kV/cm

Ebias = 0 kV/cm

Reflected magnitude (dB)

−15

−10

−5 0

Frequency f (GHz)

5 5,2 5,4 5,6 5,8 6

(b)

Phase standard deviation (°)

25 30 35 40 45

Frequency f (GHz)

5 5,2 5,4 5,6 5,8 6

hf = 0.9 µm hf = 1 µm hf = 1.1 µm

Phase standard deviation (°)

25 30 35 40 45

Scapa = 6 µm Scapa = 7 µm Scapa = 8 µm

(c)

Figure 5. Simulated reflection coefficient (phase and magnitude) of the proposed cell for different gap widths of the active capacitorScapa(a) and different thicknesses of the material of the active capacitorhf (b). Phase standard deviation computed with 4 chosen states for different width of the gapScapa(top) and for different thicknesseshf (bottom).

value is less than 32° within a 220 MHz band (Fig. 3) which corresponds to a 1.70 bit phase-shifter.

As the cell is not symmetrical (due to the biasing circuit), the cross polarization has also been verified. It remains below

−28 dB in the considered band (Fig. 4), which indicates that the addition of the biasing circuit does not extensively disturb the radiation.

V. S

The sensitivity to the physical dimensions of the cell has been investigated in simulation. The most critical parameters are the gap between the two electrodes of the active capacitor S

and the thickness of the ferroelectric material h

.

The value of the gap has been varied in simulation from 6 µm to 8 µm and the results are presented Fig. 5a. This corresponds to the uncertainty of the dimensions resulting from the etching process of copper. A variation of the gap introduces a frequency shift; this is logical because the gap controls the capacitance which loads the central slot. The thickness has also been varied from 0.9 µm to 1.1 µm and the results are reported Fig. 5b. The influence on the phase response is quite similar; indeed, this thickness controls the contribution of the active material to the effective permittivity in the capacitor and thus the capacitance itself. For the different variations of both parameters, the accessible phase range stays close to 270°.

To complete the sensitivity study, the phase standard de- viation has been computed for the different values of h

and S

, the results are shown Fig. 5c. The uncertainty is responsible for a slight variation of the bandwidth and of the central frequency.

P ha se -s hi ft ( °)

0 50 100 150 200 250 300

Applied bias eld E

(kV/cm)

0 100 200 300 400

Simulation Measurement

Figure 6. Simulated and measured phase-shifts provided by the proposed cell as a function of the bias field applied on the material.

VI. M

The proposed cell has been measured in a 35×35 mm

square metallic waveguide fed with a standard WR137 rect- angular waveguide [7]. The reference plane is set on top of the alumina substrate and three offset short-circuits are used for calibration [13].

The simulated and measured phase-shift values as a func-

tion of the applied bias field are reported in Fig. 6. The

maximum phase-shift and the overall phase range obtained

in measurement are close to the simulated ones. This shows

that the ferroelectric material is promising to realize reconfig-

urable cells. Nevertheless, the evolution with the applied bias

field is sharper in measurement. This difference comes with

higher losses than expected. The conductivity of the copper deposited by sputtering has been measured and the value is σ = 2.10

S.m

, which is much smaller than the bulk value and could explain these differences.

VII. C

An active triple-slot phase-shifting cell is presented. The use of multiple-resonators reduces the losses (less than 6 dB in simulation) and increases the bandwidth of the cell as com- pared to a single-slot topology. The cell provides a 270° phase range with a 60% tunability of the material (corresponding to a maximum applied bias field of 400 kV/cm in the case of the used BST thin film). This allows to easily derive 4 uniformly- distributed phase states. The effect of the biasing lines has also been assessed and the impact on the cross-polarization is negligible.

The sensitivity of the cell to the gap between the plates of the ferroelectric capacitor and to the thickness of the active material has been evaluated in simulation and is moderate.

The maximum phase-shifts obtained in measurement and in simulation are very close to one another. Work is underway to increase the conductivity of the deposited copper and thus to obtain a smoother variation of the phase-shift with the applied electric field.

R

[1] J. Huang and J. A. Encinar,Reflectarray Antennas. John Wiley, 2008.

[2] S. V. Hum, M. Okoniewski, and R. J. Davies,“Modeling and design of electronically tunable reflectarrays,”IEEE Trans. Antennas Propag., vol. 55, no. 8, pp. 2200–2210, 2007.

[3] J. Perruisseau-Carrier and A. Skriverviky, “Monolithic MEMS-based reflectarray cell digitally reconfigurable over a 360° phase range,”IEEE Antennas Wireless Propag. Lett., vol. 7, pp. 138–141, 2008.

[4] G. Perez-Palomino, P. Baine, R. Dickie, M. Bain, J. Encinar, R. Cahill, M. Barba, and G. Toso,“Design and experimental validation of liquid crystal-based reconfigurable reflectarray elements with improved band- width in F-band,”IEEE Trans. Antennas Propag., vol. 61, no. 4, pp.

1704–1713, 2013.

[5] Y. Shen, S. Ebadi, P. Wahid, and X. Gong,“Tunable reflectarray unit cell element using BST technology,” inIEEE Radio and Wireless Symposium (RWS), 2012, pp. 43–46.

[6] J. Encinar, M. Arrebola, M. Dejus, and C. Jouve,“Design of a 1-metre reflectarray for dbs application with 15 % bandwidth,” in6^thEuropean Conference on Antennas and Propagation (EUCAP), 2006, pp. 1–5.

[7] T. Makdissy, R. Gillard, E. Fourn, E. Girard, and H. Legay,“Triple-slot phase-shifting cell loaded with one variable capacitance for reflectarray applications,” Electronics Letters, vol. 48, no. 21, pp. 426–427, Oct.

2012.

[8] K. Nadaud, R. Gillard, E. Fourn, H. Gundel, and C. Borderon,“A simple Phase-Shifting cell for reflectarray using a slot loaded with a ferroelectric capacitor,” in Loughborough Antennas and Propagation Conference, 2014, pp. 214–217.

[9] C. Borderon, D. Averty, R. Seveno, and H. W. Gundel,“Influence of the morphology of barium strontium titanate thin films on the ferroelectric and dielectric properties,”Integrated Ferroelectrics, vol. 97, no. 1, pp.

12–19, 2008.

[10] K. Nadaud, C. Borderon, R. Renoud, and H. W. Gundel, “Effect of manganese doping of BaSrTiO3on diffusion and domain wall pinning,”

Submitted to the Journal of Applied Physics, 2014.

[11] K. Nadaud, H. W. Gundel, R. Gillard, C. Borderon, and E. Fourn,“A new method of dielectric characterization in the microwave range for high-k ferroelectric thin films,” inProc. IEEE International Symposium on the Applications of Ferroelectric, 2013, pp. 9–12.

[12] R. Pereira, R. Gillard, R. Sauleau, P. Potier, T. Dousset, and X. Delestre,

“Dual linearly-polarized unit-cells with nearly 2-bit resolution for reflec- tarray applications in x-band,”IEEE Trans. Antennas Propag., vol. 60, no. 12, pp. 6042–6048, 2012.

[13] G. J. Scalzi, A. J. J. Slobodnik, and G. A. Roberts,“Network analyzer calibration using offset shorts,” IEEE Trans. Microw. Theory Tech., vol. 36, no. 6, pp. 1097–1100, 1988.